Treffer: The method of complex analysis of differentiation of phonostatistical structures of english styles

The method of complex analysis of differentiation of phonostatistical structures of English styles has been developed. The method is based on a combination of the two statistical criteria of hypothesis verification for sample homogeneity: the Student's t-test and the Kolmogorov-Smirnov test. The combination of the criteria secures higher precision of style differentiation. On the basis of the developed method, the statistical model of determining the degree of author's style factor effect has been built. The model enables improving accuracy of style and authorship attribution. The developed method and model have been coded on the Java programming language. In the program, the POST requests of two types such as /process i/process/transcription have been developed. The first request is used when the transcription variant of the researched text is not available, the second – when it is available. The second request makes it possible to reduce the program operating time. An open, cross platform, inbuilt data base H2, written on the programming language Java has been used. The data base H2 supports the SQL and is well integrated with the used framework Spring Boot and doesn't need any additional installations. The reply from the transcription transformation site is stored in the data structure HashMap, built on the principle key-meaning and allows avoiding copies. The greater amount of data is processed, the fewer requests to the Internet are made. This secures program independence and autonomy. The program consumes little time. For testing the program, the texts from the publicist style have been chosen ("Freedom Paper", papers by S. Logan and D. Webster). The essential differences have been established in the nasal, dorsal and velar phoneme groups by the Student's t-test. The essential differences have been established in all eight phoneme groups by the Kolmogorov-Smirnov's test. The statistical model of author-differentiating capability for the fri